Peer-reviewed EPA software, ProUCL has been developed to address statistical and data analytic needs of environmental projects originating from the various CERCLA and RCRA site investigations. ProUCL can be downloaded from the EPA website: http://www.epa.gov/nerlesd1/tsc/TSC_form.htm
ProUCL Webinar Part I will provide basic information regarding: ProUCL logistics and computer requirements; data file creation and management for full data sets without nondetect observations and for left-censored data sets consisting of nondetects (NDs); analysis of multiple contaminants simultaneously; analysis of data with multiple groups; saving graphs and output sheets generated by ProUCL.
The focus of ProUCL Webinar Part I will be to make participants familiar with graphical and statistical methods available in ProUCL to: identify outliers; perform goodness-of-fit tests for normal, lognormal, and gamma distributions; compute DQOs based minimum sample sizes needed to address project objectives; and compute 95% upper confidence limits to estimate exposure point concentration (EPC) terms. Participants will learn how to interpret statistical results produced by ProUCL and derive conclusions based upon those results.
Data sets consisting of NDs will be used to illustrate statistical methods for left-censored data sets. Real data sets from participants' projects can also be considered to address statistical issues often encountered in environmental applications. The participants may send their data sets in advance.
Who Should Attend
This webinar will be useful for regulators, environmental scientists and practitioners, site owners/operators, consultants and contractors who use or want to use statistics to do their work.
Outline (all times EST)
- 1:00-1:10 (Felicia Barnett)
Welcome and Introduction
- 1:10-1:20 (Robert Maichle)
ProUCL4.1 logistics, computer hardware and software requirements
- 1:20-1:45 (Anita Singh)
ProUCL Input/ Output Operations - Data file creation for data sets with and without nondetect values; analysis of multiple contaminants; analysis of data sets with multiple groups; saving graphs and output sheets generated by ProUCL
- 1:45-2:05 (Anita Singh)
Determining DQOs based minimum sample sizes needed to estimate mean; to perform single sample and two sample hypothesis tests; and acceptance sampling
- 2:05-2:35 (Anita Singh)
Graphical capabilities of ProUCL4.1 for data sets with and without nondetect values including Box plots and Q-Q plots to identify outliers and compare multiple groups; Time series plots to identify trends in data collected over a period of time
- 2:45-3:15 (Anita Singh)
Dixon and Rosen outlier tests; Goodness-of-Fit (GOF) tests for Normal, Lognormal, and Gamma distributions for data sets with and without nondetect values
- 3:15-3:45 (Anita Singh)
Computation of 95% Upper Confidence Limit to estimate exposure point concentration (EPC) term for uncensored full data sets without nondetect values